Image decoding device, and image encoding device

Abstract

Included is an inter-layer merge candidate derivation unit 1464 B that derives motion information decoded in an area on a reference layer corresponding to an area which is located at a lower right part of a target PU and for which decoding has not been performed, in the target PU of a target layer, as a merge candidate.

The invention claimed is: 
     
       1. An image decoding device which decodes hierarchically encoded data, restores motion information of a decoding target layer, and restores an image of the target layer by using a predicted image generated by motion compensation prediction based on the restored motion information, comprising:
 a merge candidate derivation circuit configured to derive a plurality of merge candidates in a target prediction unit for generating the predicted image on the target layer, 
 wherein the merge candidate derivation circuit includes, as one of individual merge candidate derivation means, an inter-layer candidates derivation circuit configured to derive a merge candidate called an inter-layer merge candidate, by using a motion information which is decoded and associated with a predetermined region on a reference layer, 
 wherein the predetermined region is a position corresponding to a lower right position derived by adding an offset, having a lower right direction, to a pixel position on a reference layer calculated by referring to a top-left pixel position of the target prediction unit, 
 wherein the merge candidate derivation circuit further includes, as one of the individual merge candidate derivation means, a spatial merge candidate derivation circuit configured to derive a merge candidate called a spatial merge candidate, by using motion information of positions of an upper right part, a lower left part, and an upper left part of the target prediction unit, 
 wherein a plurality of merge candidates derived by the merge candidate derivation circuit include the inter-layer merge candidate and the spatial merge candidate, 
 wherein a recording position of the motion information, which is decoded and associated to the predetermined position, is represented by ((xC0//N)*N, (yC0//N)*N), where (xC0, yC0) is the predetermined position and N is positive integer value. 
 
     
     
       2. The image decoding device according to  claim 1 ,
 wherein the inter-layer candidate derivation circuit derives an inter-layer merge candidate by referring motion information which is decoded and associated with a first position on a reference layer when the first position is available, 
 and the inter-layer candidate derivation means derives the inter-layer merge candidate by referring motion information which is decoded and associated with a second position, that is located on the upper left of the first position, when the first position is not available. 
 
     
     
       3. The image decoding device according to  claim 1 , further comprising:
 a time motion information candidate derivation circuit configured to derive the merge candidate based on the motion information decoded in a picture other than a picture to which the target prediction unit belongs, 
 wherein the merge candidate derivation circuit execute either an inter-layer merge candidate derivation process, which is conducted by the inter-layer candidate derivation means, or a merge candidate derivation process, which is conducted by the time motion information candidate derivation means. 
 
     
     
       4. An image encoding device which generates a prediction residual of a target layer by subtracting a predicted image which is generated by motion compensation prediction based on motion information which is derived in the target layer which is an encoding target, from an input original image of the target layer, in a case of generating hierarchically encoded data, comprising:
 a merge candidate derivation circuit configured to derive a plurality of merge candidates in a target prediction unit for generating the predicted image on the target layer, 
 wherein the merge candidate derivation circuit includes, as one of individual merge candidate derivation means, an inter-layer candidates derivation circuit configured to derive a merge candidates called an inter-layer merge candidate, by using a motion information which is decoded and associated with a predetermined region on a reference layer, 
 wherein the predetermined region is a position corresponding to a lower right position derived by adding an offset, having a lower right direction, to a pixel position on a reference layer calculated by referring to a top-left pixel position of the target prediction unit, 
 wherein the merge candidate derivation circuit further includes, as one of the individual merge candidate derivation means, a spatial merge candidate derivation circuit configured to derive a merge candidate called a spatial merge candidate, by using motion information of positions of an upper right part, a lower left part, and an upper left part of the target prediction unit, 
 wherein a plurality of merge candidates derived by the merge candidate derivation circuit include the inter-layer merge candidate and the spatial merge candidate, 
 wherein a recording position of the motion information, which is decoded and associated to the predetermined position, is represented by ((xC0//N)*N, (yC0//N)*N), where (xC0, yC0) is the predetermined position and N is positive integer value.